no. 2.] PSYCHOLOGICAL EXAMINING IN THE UNITED STATES AEMY. 345 



to take the beta examination. As for the second criticism, it is true that the intercorrelations 

 of the alpha tests are slightly higher than the tests of examination a. It is probably also true 

 that the number of alpha tests could be reduced to four or five without serious loss. On the 

 other hand, this is not an argument against the elimination of the two tests which had nothing 

 to commend them except low correlations with the other tests. As shown on page 452, exami- 

 nation a correlates as well with officers' estimates when tests 2 and 10 are not counted in the 

 total as when they are. 



Doubtless alpha could be improved. Many of its items are unsatisfactory and for no test 

 are they arranged in order of difficulty. Some of the tests are rather too hard for average 

 enlisted men. It is possible that 12 or 14 briefer serial tests, requiring in all about the same 

 time as the present 8, would be more effective. It is also possible that three or four of the best 

 alpha tests combined with three or four of the best beta tests would be appreciably better than 

 either alpha or beta alone. It is practically certain that alpha can not be greatly improved by 

 extending it along similar lines or by giving a second form of the test at the same sitting. It is 

 even doubtful whether a much better measure would be secured by averaging the results of 

 two or three repetitions of the tests at different sittings. It is also unlikely that any great 

 improvement will be accomplished by changes in time allowance or the method of scoring. It 

 is probable that improvement should rather bo sought in the more effective combinations 

 of alpha tests with other tests of somewhat different nature, perhaps with "o mni bus" tests 

 involving more frequent change of problem. 

 121435°— 21 23 



